Ultra-Low-Field Portable MRI on Patients Undergoing ECMO

Extracorporeal membrane oxygenation (ECMO) increases the risk of acute brain injury (ABI), including intracerebral hemorrhage (ICH), ischemic brain injury, and hypoxic brain injury.¹ Neurologic complication is also a leading cause of death and decannulation in those on venoarterial (VA) ECMO.² However, due to sedation and often paralysis, obtaining a proper neurologic examination on patients undergoing ECMO is difficult or impossible. Multimodal neuromonitoring has been shown to improve patient outcomes, but many patients on ECMO do not have CT scans due to the dangers of moving a patient on ECMO.³ Further, neurologic complications, including ischemic brain injuries, are often missed by current neuromonitoring modalities such as transcranial Doppler, near-infrared spectroscopy, EEG, and CT. Kang et al attempted to solve this clinical deficiency by studying the safety of ultra-low-field portable MRI on patients on ECMO.⁴

This prospective observational study was conducted in the intensive care units (ICUs) at two large U.S. tertiary care centers. Patients were enrolled between August 2021 and November 2022 and included those older than 18 years undergoing ECMO, including patients with Swan-Ganz catheters or intra-aortic balloon pumps (IABPs). Excluded patients were those weighing more than 200 kg (440 lb) or having any metal implants such as pacemakers, defibrillators, nerve stimulators, cochlear implants, MRI-incompatible surgical hardware, artificial heart valves, or other irremovable ferromagnetic devices. A total of 53 patients were enrolled, three of whom were excluded due to severe facial or scalp swelling. Of the 50 patients included, 34 were on VA ECMO, with 11 of those centrally cannulated, and 16 were on venovenous ECMO. Median age was 58 and median Glasgow Coma Scale score was 6. 

Point-of-care MRIs were performed in each patient’s room by a trained operator. After the patient was aligned with the system, a team of four to six staff moved the patient into the scanner. All ICU equipment was kept outside the 5-gauss line, and the patient was kept as flat as possible during the examination. All images were read by a neuroradiologist at each site.

All 50 patients were imaged successfully. Three adverse events were recorded, one of which was determined to be serious and related to the MRI—an IABP malfunctioned due to interference of the ECG leads. The IABP was then switched to pressure trigger mode, and the examination was resumed successfully. Eight patients were found to have imaging artifacts but only one was determined to affect the ability to diagnose ABI. This artifact was seen only on a diffusion-weighted sequence and the presence of infarct was uncertain.

ABI was found in 22 (44%) of the 50 patients, with ischemic stroke identified most frequently in 18 (36%) patients, followed by ICH (n=3, 6%) and hypoxic-ischemic brain injury (n=2, 4%). Thirty-six patients had both MRI and CT, but only 18 of these patients had them within 24 hours of each other. Of these 18 patients, 9 were found to have an ABI with a total of 10 events—eight ischemic injuries and two ICHs. The MRI was able to show all eight of these ischemic injuries whereas the CT showed only half of them. Conversely, of the two ICHs, the MRI showed only one of them, while CT showed both.

This study’s limitations were mostly related to its limited scope and procedural difficulties. Of the 50 patients who had an MRI, only 36 also had a CT, limiting the ability to compare the two modalities. A major limitation was that, of the 36 patients, only 18 had both studies within 24 hours of each other. This time difference increased the likelihood of any observed ABI occurring during the period between CT and MRI.

Despite these limitations, the primary outcome of determining the feasibility of performing an ultra-low-field MRI on an ECMO patient was accomplished. Neuromonitoring is a complicated part of ECMO patient care with no nationally standardized protocol to guide management. This study showed that MRI can be performed on these patients, and it significantly advances neuromonitoring capability.


References

1. Lorusso R, Barili F, di Mauro MD, et al. In-hospital neurologic complications in adult patients undergoing venoarterial extracorporeal membrane oxygenation: results from the Extracorporeal Life Support Organization Registry. Crit Care Med. 2016 Oct;44(10):e964-e972.
2. Belohlavek J, Yannopoulos D, Smalcova J, et al. Intraarrest transport, extracorporeal cardiopulmonary resuscitation, and early invasive management in refractory out-of-hospital cardiac arrest: an individual patient data pooled analysis of two randomised trials. EClinicalMedicine. 2023 May 5;59:101988.
3. Ong CS, Etchill E, Dong J, et al. Neuromonitoring detects brain injury in patients receiving extracorporeal membrane oxygenation support. J Thorac Cardiovasc Surg. 2023 Jun;165(6):2104-2110.e1.
4. Kang JK, Etchill E, Verdi K, et al. Ultra-low-field portable MRI and extracorporeal membrane oxygenation: preclinical safety testing. Crit Care Explor. 2024 Oct 18;6(10):e1169.